Only 7 percent of companies create value with AI. Operations are the barrier, not technology.
MegazoneCloud CEO Yeom Dong-hoon said on April 2 that most companies fail to extract value from AI because they lack the operational framework to manage governance, security and compliance at scale. He made the comments at ICON 2026, the company's partner conference in Seoul.
The gap is stark. While companies spend heavily on AI models and infrastructure, only 7 percent actually generate measurable business value. The problem isn't the technology itself.
Operations determine success, not model quality
Hwang In-cheol, MegazoneCloud's chief revenue officer, said a well-operated structure matters far more than a well-built model. Companies need to embed compliance requirements and regulations into their systems from the design phase, not bolt them on later.
This reframes how executives should think about AI investment. The expensive part isn't the model. It's the operational infrastructure needed to run it safely and reliably across the organization.
The five standards for AI operations
MegazoneCloud introduced its Enterprise TRUST Layer strategy, which sets five standards for managing AI environments:
- Traceability - tracking decisions and data flows
- Regulation - embedding compliance rules
- User Access - controlling who can do what
- Standardization - consistent processes across teams
- Tooling - integrated platforms for management
The company also unveiled AIR Studio V2, an enterprise AI operating system that integrates model management, data governance, orchestration and control into a single platform.
Scaling beyond proof of concept remains the real challenge
Intel and Articul8, both presenting at the conference, identified the same bottleneck: companies succeed at AI pilots but struggle to scale into production.
Jason Tan, who oversees Intel's Asia-Pacific AWS business, said the bigger challenge isn't adopting AI - it's running it stably at scale after the proof of concept ends. Edward Kong from Articul8 said companies face practical barriers in moving from experimentation to real work environments, including security, data reliability and operational stability.
This explains why technology vendors alone can't solve the problem. Companies need systems thinking, not just better tools.
Security and visibility drive structural change
A panel discussion with Korean Air and cloud security company Wiz showed how operations improvements create business value. Kim Hyo-jong, a team leader at Korean Air's information security office, said the airline's adoption of Wiz's system felt like "a fog had lifted" - the security team could finally see which risks actually mattered in a multi-cloud environment instead of drowning in alerts.
Matt Zwolenski, senior director at Wiz, said the key isn't adding more security tools. It's securing visibility to identify which risks matter in complex environments and managing them structurally. Different teams need to build workflows based on a common language.
Wi Soo-young, head of MegazoneCloud's security unit, said security is not a defensive technology but an operational design that prevents business interruption. Better visibility also improves how teams collaborate across the organization.
The path forward for executives
For executives evaluating AI investments, the lesson is clear: budget for operations, not just models. Governance frameworks, compliance systems and integrated management platforms determine whether AI creates value or becomes expensive technical debt.
Learn more about AI for Executives & Strategy and AI for Operations to understand how to structure AI programs for sustainable value.
Your membership also unlocks:
